5/18/2021
degenerate
EvgaUser2711201
degenerate
EvgaUser2711201
Ppl have been telling me they see over 500w on pr with 520w but I guess I’ve got a dud

You'll need to push voltages in order to see that kind of power draw in a PR run.
 
What's your max OC score using stock out of the box settings (LN2 bios, max power/voltage sliders)?


https://www.3dmark.com/pr/1028918 
This was my best before the update that messed with the scoring. 
ln2, maxed powerlimit and core slider. 
 
Is there a kp hydro copper ln2 bios? Havent seen it on techpower up.
I need to format pc, ill report back tomorrow with my findings.  
 
 
 
 


You're breaking 15k with no changes to voltage my man, that's not a dud.
 
No there's only the one set of bios' for the Kingpin cards. Normal, OC, LN2 and XOC.
5/18/2021
Dabadger84
emmett
CptT95
Lower core temps results in lower mem temps on backside chips
Plus lower core temps means less power uses, and more stable clocks



I am just hitting such a hard wall trying to go over 2205 starting at 8C with portable AC blasting through radiator in a very cold room that I do not think the block
will do anything for me. I gained about 100-110 MHZ on my Strix after installing block, but that was hitting 60-70 on the air cooler.




The big thing with the block is how much it will lower your other temperatures, thus decreasing PCB heatsoak that effects temperatures even further.  Memory would be running much cooler, VRMs will be about the same (unfortunately, but they don't really run hot to begin with).  The core temp itself for me didn't go down a ton, but I was also dealing with a "base" starting temp in the low 50s for most tests/games, so it can only GO so much lower at that point.
If you're having issues getting over 2205 it sounds like your card is probably one of those voltage-happy ones that will clock pretty well under LN2 but at "stock" voltages, can't OC too great.  I know mine is pretty similar, for average boost clock, I can't consistently get higher than 2205-2220... but I've also "only" went up to 1.14375V NVVDD with a 1.1V MSVDD.  That allowed me to run +165-180, mostly stable, for Port Royal specifically.  The other benchmarks I have to step it down a bit or they crash almost all the time.
 
Ideally I'd like to hit at least 2205 stable enough to use, but meh.
 
I'm gonna try to run Metro Exodus Enhanced Edition tonight/this morning with some OC settings & see if it crashes or not, and if it doesn't, what the power draw is like.  With my Undervolt/OC to 2100MHz @ 1025mV, it was getting up to around 450W draw, without VSync on, so it's definitely a power hungry game... but man is it purdy.
 
EvgaUser2711201
Ppl have been telling me they see over 500w on pr with 520w but I guess I’ve got a dud



I think at stock on Port Royal, and even with a "regular" OC, I'm only seeing around 480-500W.  I have to bump voltages in Classified Tool to get consistently over 500W draw.  Highest I've see was on the card's OLED set to power readout during my personal best score runs & it got as high as 551W, and that's pushing 1.14375V core with the memory at 1.4125V +1400 offset... pretty sure I was hitting power limit at times during that, and prolonged testing with those settings saw the typical PCB heatsoak symptoms of the score going down for seemingly no reason, clocks dropping a bit and whatnot on the core.
5/19/2021
Dabadger84
https://www.3dmark.com/3dm/61871730 3DMark is fixed, I repeat, 3DMark is fixed.
 
That's with no classified Tool & Afterburner/HWInfo open in the background.
5/19/2021
Jolly9612
Was that 5GHz 5950X all core or just single core?
Dabadger84
https://www.3dmark.com/3dm/61871730 3DMark is fixed, I repeat, 3DMark is fixed.
 
That's with no classified Tool & Afterburner/HWInfo open in the background.
5/19/2021
degenerate
Dabadger84
https://www.3dmark.com/3dm/61871730 3DMark is fixed, I repeat, 3DMark is fixed.
 
That's with no classified Tool & Afterburner/HWInfo open in the background.


Yeah I scored 15.2k no classified or the usual performance enhancing "tricks". Good stuff!
 
https://www.3dmark.com/3dm/61872397
5/19/2021
andreibondrea
Who has a 3090 to post results? :D
5/19/2021
fragility_V1
Just got a 15,757 pulling out all the stops with the classy tool.  New card to me, had to RMA my original KP.  Put the HC block right on the RMA replacement, so no good point of comparison to stock cooling, but is performing a little better than my previous card.  Now if it just wasn't so warm all the time in SoCal.
 
https://www.3dmark.com/pr/1045077
 
Should be somewhere in the 70s on single card PR, but drivers haven't been approved yet. 
5/19/2021
Dabadger84
Jolly9612
Was that 5GHz 5950X all core or just single core?
Dabadger84
https://www.3dmark.com/3dm/61871730 3DMark is fixed, I repeat, 3DMark is fixed.
 
That's with no classified Tool & Afterburner/HWInfo open in the background.




CPU was 100% stock for that run, only changes were FCLK to 1900MHz & RAM @ 3800MHz with DOCP timings for the kit (14-15-15) with some of them set manually by me (tRFC is 312 on the runs, I think).
 
Hit a new personal best, good enough for #4 on the 5950X/Single 3090 "leaderboard"/HoF:
 

 
https://www.3dmark.com/pr/1045048 and that's with the same settings as my previous 15668 run too:
 

5/19/2021
degenerate
fragility_V1
Just got a 15,757 pulling out all the stops with the classy tool.  New card to me, had to RMA my original KP.  Put the HC block right on the RMA replacement, so no good point of comparison to stock cooling, but is performing a little better than my previous card.  Now if it just wasn't so warm all the time in SoCal.
 
https://www.3dmark.com/pr/1045077
 
Should be somewhere in the 70s on single card PR, but drivers haven't been approved yet. 



Dabadger84
Jolly9612
Was that 5GHz 5950X all core or just single core?
Dabadger84
https://www.3dmark.com/3dm/61871730 3DMark is fixed, I repeat, 3DMark is fixed.

That's with no classified Tool & Afterburner/HWInfo open in the background.




CPU was 100% stock for that run, only changes were FCLK to 1900MHz & RAM @ 3800MHz with DOCP timings for the kit (14-15-15) with some of them set manually by me (tRFC is 312 on the runs, I think).
 
Hit a new personal best, good enough for #4 on the 5950X/Single 3090 "leaderboard"/HoF:
 
Spoiler


 
https://www.3dmark.com/pr/1045048 and that's with the same settings as my previous 15668 run too:
 

close




5/19/2021
magillaman
So here's the difference between running "stock" and using "tools" for those HOF scores.
 
Here's PR after 2hrs of Cyberpunk and Mass Effect (10 min cool down), while using my daily driver settings: Default voltages, no Classy tool, Afterburner curve at 2205 @ 1.068v, memory +1200. 
 
PR 15,751

 
Compared to this run, which required:
Ambient temp in the room lowered to 15c, Classified tool set to 1.16875 NVVD, Afterburner set to 2310 @ 1.1v, mem +1750, right after system start up so before anything warmed up
 
PR 16,312
 
Just wanted to put the performance gains in perspective.
 
It's fun pushing the cards once in a while, and with the KP you're sort of "supposed to" as it's an enthusiast card. Getting high scores is tricky, but you learn about the card and it helps find the "sweet spot" for every day use. 
 
Side note: I'm actually impressed by this waterblock.

Use My Existing Forum Account

Use My Social Media Account

loading